My first spectrum analyzer - Second hand or new USB based

[TheUnnamedNewbie]

So I am planning to get myself a spectrum analyzer. I have no real idea about budget yet, but I suspect somewhere between 2-3k EUR is the limit (though I have no requirement to spend it all). I'm not a HAM, my main goal is to do fun experiments and repairs of gear, and right now I don't have a spectrum analyzer to do repairs on anything above 50 MHz.

As I see it, I have three (and a half) options I can look at: either, I start with something like a HP 8594E or 8595E. 2.9 or 6.5 GHz, without a tracking generator these go for below 1k USD/1.5k USD sometimes. Not quite the frequency span I hope to work with in the long run, but they could function as a good first spectrum analyzer and allow me to wait for a great deal on a more modern, higher frequency (20+ GHz) analyzer.

The other option is to go ''big'' straight away and get something like a 8563E. 26.5 GHz, seen sometimes for about 2.5k USD, and that should keep me covered for a long time and allow me to do many repairs as well.

The final is to go for a new spectrum analyzer in the form of a Signal Hound USB-SA124B. 12 GHz, 2-2.5K USD, but this would be a new device instead of an old HP spectrum analyzer.

As I am new to the world of spectrum analyzers I don't know what features/abilities different models have, so I find it hard to compare different models. Are there any other options I should be considering? Apart from the fact that a second-hand HP one will be a standalone unit, does it have any performance benefits over a new Signal Hound? Any input on this is welcome.

[DaJMasta]

Do you actually have a use for the high bandwidth options?  Sure 20+GHz sounds really impressive, but you need much more expensive cables, adapters, probes, etc. to actually use much of it.  If you stick to <15GHz, at the very least, you can probably get an N connector (or maybe a 3.5mm connector) and have an easier time connecting to more common lower frequency stuff.

Regardless of that, the expense often comes with bandwidth, so think on how much you really want to be able to see, and see if your budget can stretch farther without the extra bandwidth - maybe more options, maybe a more modern SA, maybe some accessories you were looking at.

Once you get beyond 6GHz or so, you really don't run into all that many signals floating around in the air in a home lab, and you'll only run into most of those if you're also generating the signal (and a 10GHz signal generator is not a cheap thing), or you're working on high frequency RF gear.

Also worth considering form factor and connector options, if you've got a computer already (and you probably do), a USB SA has a tiny space requirement and can be a great option, but maybe you want it to be standalone (other things using the computer, limited screen space, whatever), maybe you want a direct IF output or I/Q inputs or outputs, and with most USB SAs, they're not available.  If physical space is a concern, you have the option of the more modern SAs that aren't deep, or going after a portable SA like an FSH or SpectrumMaster.  There's pros and cons to all sorts of stuff, but until you know the kind of stuff you want to do and have some experience using them, it will be difficult to decide when you have a very broad selection to choose from.

[nctnico]

Don't forget to look at the spectrum analysers from Advantest. These are very good and don't come with the HP name premium.

[G0HZU]

As I see it, I have three (and a half) options I can look at: either, I start with something like a HP 8594E or 8595E. 2.9 or 6.5 GHz, without a tracking generator these go for below 1k USD/1.5k USD sometimes.

In terms of what options are available to you, a lot has happened in the last 10 years or so and so you can afford to be very choosy and buy something decent for your budget even if your future needs are fairly basic. For this reason I think old dogs like the HP859x series should be low on your shopping list in 2018.

The HP859x series of spectrum analysers can be best summed up as having early 1970s technology wrapped up in a late 1980s portable package. They have limited log range, poor phase noise, limited resolution bandwidth on some models and an awful display. They would be fine for basic tasks like checking spurious on a ham radio or for casual spectrum monitoring but these analysers were dogs even when they were brand new 25-30 years ago. The E versions address some of the limitations but they are just an attempt at polishing an ageing turd.

In 2018 as a home/hobby user you really need to be comparing against the option of buying a new Siglent SSA3021X 2.1GHz (hacked to 3.2GHz?) spectrum analyser with its modern digital IF and modern connectivity and decent display. For many hobby/home users, this seems to be a popular choice with plenty of happy users. So if you stray away from this analyser you need to try and justify why. Possible reasons would be the need to capture/process data on a PC or because you want to do fairly serious RF design that requires a bit more dynamic range  and lower phase noise and lower overall measurement uncertainty than the Siglent can offer.

[TheUnnamedNewbie]

In 2018 as a home/hobby user you really need to be comparing against the option of buying a new Siglent SSA3021X 2.1GHz (hacked to 3.2GHz?) spectrum analyser with its modern digital IF and modern connectivity and decent display. For many hobby/home users, this seems to be a popular choice with plenty of happy users. So if you stray away from this analyser you need to try and justify why. Possible reasons would be the need to capture/process data on a PC or because you want to do fairly serious RF design that requires a bit more dynamic range  and lower phase noise and lower overall measurement uncertainty than the Siglent can offer.

Perhaps you are correct and I was wrong to dismiss the Siglent so soon. The only thing that bothers me is the bandwidth, 3.2 GHz really does seem a bit on the low side - my idea is to get the first analyzer so I can repair others in the future, and getting only 3.2 GHz does limit me in that respect somewhat. But maybe, for a first analyzer to get started, it can be a good purchase.

[Mechatrommer]

if you can afford descent price like stated 3K EUR... there are few rules according to me (and i am good )

1) dont buy USB based, that will be PC dependent. built in monitor will be golden. handheld unit if you think you'll be working outdoor.
2) get greater than 5.8GHz, why? current hobbiest rc system (wireless video, receiver transmitter) has reach that level. so if you get 1GHz SA and later buy toy stuff with 5.8GHz radio in it, you are out.
3) get brand name if possible, like HP/Rohde etc. if price is out bound, get other brand like Advantest, Rigol Siglent etc.
4) if there is room in the budget, get one with tracking generator, this may help you characterize your circuit respond.
5) if buy 2nd hand, read very carefull about item condition, go ask the seller if necessary and get enough picture.

and many more which i forgot... anyway, i have a keen eye on used HP SA (anything beyond and 6GHz BW). but they are out of my budget, so i have to buy out of calibration and near death Advantest. i have to open it up sometime to repair some of it failure in calibration process. i'm guessing some of the opamp inside are in twilight condition. ymmv.

[nctnico]

In 2018 as a home/hobby user you really need to be comparing against the option of buying a new Siglent SSA3021X 2.1GHz (hacked to 3.2GHz?) spectrum analyser with its modern digital IF and modern connectivity and decent display. For many hobby/home users, this seems to be a popular choice with plenty of happy users. So if you stray away from this analyser you need to try and justify why. Possible reasons would be the need to capture/process data on a PC or because you want to do fairly serious RF design that requires a bit more dynamic range  and lower phase noise and lower overall measurement uncertainty than the Siglent can offer.

Perhaps you are correct and I was wrong to dismiss the Siglent so soon. The only thing that bothers me is the bandwidth, 3.2 GHz really does seem a bit on the low side - my idea is to get the first analyzer so I can repair others in the future, and getting only 3.2 GHz does limit me in that respect somewhat. But maybe, for a first analyzer to get started, it can be a good purchase.

Actually the Siglent is a pretty good option. I seriously considered buying one but I spotted a nice deal on Ebay for an Advantest spectrum analyser (which needed a little bit of TLC). Don't get too hung up on bandwidth. With every GHz extra the prices of the cables double.

[Mechatrommer]

Don't get too hung up on bandwidth. With every GHz extra the prices of the cables double.

if we look, patient and lucky, there are cables that can be had for cheap. i got a few rated to 24GHz (SMA connectors) for use with my 6-8GHz instruments (cant verify flatness beyond that though, within my BW its acceptable). but the fact that more BW = more cable price is undeniable, so get the BW that is reasonable for you (not nctnico). ymmv.